Saving and loading a model in Pytorch?

Don’t we need to store the optimizer states as well?